sansan0/trendradar

TrendRadar is an AI-powered news and trend monitoring tool that aggregates trending topics from multiple Chinese and international social platforms, filters them by keyword, runs AI analysis on the results, and pushes curated alerts to your preferred messaging channels. The core problem it solves is information overload: instead of manually checking dozens of news sites and trending lists each day, TrendRadar does it automatically and sends you only what you actually care about. How it works: the tool pulls trending content from platforms like Weibo, Zhihu, Douyin, and others via an aggregated API, as well as from any RSS feeds you configure. You define keywords and topics you want to track. The system then filters incoming items against those keywords, optionally sends them through an LLM, supporting providers like OpenAI, DeepSeek, and others, for intelligent filtering, translation, and analysis, and delivers a digest or individual alerts to your phone or computer. Notification channels include WeChat (personal and enterprise), Feishu (Lark), DingTalk, Telegram, Slack, email, ntfy, and Bark. The AI layer can do more than filter: it can translate foreign-language items, generate brief analytical summaries, assess sentiment, and forecast trends. The tool also supports MCP (Model Context Protocol) integration, which means you can connect it to AI assistants that support that protocol for natural-language querying of the monitored data. You would use TrendRadar if you need to stay on top of breaking news, track public opinion about a brand or topic, or monitor competitive landscape signals without spending hours scrolling social media. It is particularly geared toward Chinese-market monitoring. The tech stack is Python, deployable via Docker in under 30 seconds according to the README, with data stored locally or in the cloud.